GainForest: Equitable nature finance with regenerative intelligence

Transcript

I'm super excited to give this talk. Um, I even was so excited I switched my profile picture briefly from like a serious one to that cartoon one that I love. I'm here to talk about gain forest which is equal to nature finance with regenerative intelligence. That's a mouthful, but we're gonna start very clear and I'm gonna start with a sound clip. I want you to guess what animal you're gonna hear. This sound clip comes straight from the Amazon from my mouse that I recorded with my bare hands. So, listen a little bit to the rumble.

Can you hear the rumble a little bit? Can anyone guess that rumbly animal? Ha monkey. Oh my god, you took the pun away. So when I first listened to that sound, that rumble, and I was in the rainforest trying to collect that sensor, I thought, my god, not only is my mom missing my PhD ceremony, how do I explain to her that I'm getting eaten alive in this jungle right now, right? And luckily, my friend, Marina Mura, who's an indigenous scientist from the MUA tribe, she was next to me retrieving that audio sensor that has been recording exactly that clip. She was

just laughing because as correctly identified in the audience, it's just this little guy with a big big mouth doing this scary sound. It just shows me that me personally like growing up and like studying computer science in in Europe, I have no idea about any of these sounds about nature. And not only that, the world has no idea about any of these sounds of nature. In fact, science and biodiversity scientists and researchers, we have a database called GBIF, global biodiversity information facility. We stored every single observation humanity has ever done on any nature from Alexander from Humbot's first records in South America on some plants to the I naturalist picture that you just take on the phone. There's a

database with every single row and you can count them and only 1% of all observations are from the area that you just heard, right? 83% is from Europe and North America, 17% on the rest of the world, but 1% only on the rainforest. Which means if you develop any AI models, if you want to do fair payments, if you want to do any decision- making, these AI models going to be screwed. You're going to be unfair. It's going to be biased. And we also shown that in the science. But there's an there's an opportunity like marina and many of the marinas living in that area can tell us a lot about these sources, these animals, their beauty that needs to be protected, right? And so what GameForce is doing briefly is that we're building open tools for

regenerative intelligence, which means that we're trying to capture and build local first technology that integrates community knowledge but keeping community sovereignty and data sovereignty at the hands of the people that are collecting the data while giving them the advanced technological tools that are developed um to support data collection but support it in an ethical way which I'm going to tell you in a second. So we're developing AI assistance as you can see Marina here has been using the game force platform in her local comedian in Inahhabi. We implemented bio acoustic sensors around the world which is the sound that you just heard. And we did remote sensing of course with drones that communities are basically um controlling and capturing like the canopies of the forest.

But this is not about an AI talk or this is not about like a nature talk. This is a talk for the Ethereum defcon because gainforest from its early roots until now comes from the Ethereum public goods funding community. We would not exist without the first grant from Gitcoin in 2022 that gave us like our funding. we wouldn't exist from Filecoin supporting us with a big grant um that we earned in 2023 and also the subsequent Gitcoin climate rounds the cello grants um hyperserts all these beautiful ideas that our community the public goods funding committee in Ethereum which I think is one of the most beautiful ideas that web three has beautiful communities that we have given to the world we

wouldn't exist without this one but we know we all see members of the community knows we are having a really really tough year. Gitcoin closed its funding round November 2025 last year. Um, Optimism has stopped its retro funding and paused in June. This year they announced and you all have heard about Artisan that has ceased active operations and is not paying out the matching funds that were promised for so many of the amazing projects that our community has. We having a tough year. Like let's admit it, it's really uncertain how the public goods community are moving forward from this year. Definitely whatever happens next is going to be different from what happened before. But I want to have a

talk about the big three ideas that I believe our community can give to the world and leave to the world. Um and hopefully is our legacy as Audrey says a good enough ancestor to others to build on top of that. Number one, Ethereum enabled last mile payments. This is a huge humanitarian achievement. Like this is insane and we need to get really good at this. Number two, we invented, developed, experimented radically new funding mechanisms that are exciting and novel and need to be more studied. And lastly, we pioneered through governance and work committee first. Let me go through each of these points and explain why I think these are the big ideas our

committee has given to the world and the world needs to like take this on and work on towards it. Number one, Ethereum enabled last mile payments. Man, the first time I tried to pay any of the indigenous communities in the Amazon, I was calling, so I'm based in Switzerland. I was sending from my UBS bank $20 to the local community. Two weeks later, my bank calls me. What the heck is this? Is this money laundering? Like um can you send all these forms? And it cost me 20 Franks to just pay 40 Franks. It took me like two weeks. This is not an exception. This is regularity especially sending like money to really really far away regions that are uncommon for banks and in purpose. It's like usually it takes way longer. It's

often times some of these transactions never come. Ethereum solve this problem by enabling everyone who's unbanked to get a bank account. Oftentimes indigenous and local communities don't even have a paperwork and the amount of the fees are insane that allows you to do have microtransactions and it's a 12 seconds for every block instead of like days that allows us to create something called an weekly microp payment loop. Here are the most recent 47 payments that we did to 22 communities around the world that are in between $10 to $1,000. And that is no problem through the Ethereum payment rail. Many of them have never received or had a bank account in the past. They were like relying either on mobile money or on cash. And this is really a big big thing for many of them

in the world. It's a big achievement from our community. We should cherish it and we should celebrate it. The other thing is that we also have many different technologies in our circle that can empower this infrastructure. Ethereum as a money settle engine but ad protocol for example as a datacentric engine that can keep track of the receipts and this is what gainf has been building on we've been not exclusively focusing on ethereum but also also looking into nason technology can support our cause also help to catch each other out once you give them like that ability organically they are going to help each other and going to give and curate guides and so we have guides for dozens of countries from Kenya to the

Philippines from Brazil that are self-created, self-created because people care. Number two, radically new funding mechanisms. Gosh, our community has created mechanisms that economists haven't thought about for years. And in that couple of years that we had as a public goods community have pushed forward, we have created more mechanisms that I believe that the public goods academic funding committee has studied over decades. That includes impact evaluators, consistent things that are evaluating and paying out impact certificates like hyperserts, credit funding mechanism like Bitcoin and Marv.

And we're going to show you why these mechanisms will outlast whatever we're doing here. Number one, impact evalators. Impact evaluators are you can imagine like block rewards but for anything. So if you have an thing that evaluates and instantly pays you out in let's say one block which is in our case we call the bio blitz one week for whoever collects the most data you can incentivize positive feedback loops even for minimal amount of transaction costs. This is live data from all the images that we're collecting from around the world right now. Um ever since May we have collected more than almost 100,000 sound recordings, hundreds of hours from biodiversity.

15,000 species sightings have been observed by 89 people around 28 countries. And how we did it is when we launched the platform first, we had no impact vitor. And this is you can see the amount of data that has been gathered without an impacttor. Once you implement an impacttor and this is basically every week we pay out the whoever collects the most data and the most high quality data, you can see an explosion in an increase of the incentive alignment which we of course know about proof of stake. of course know about like proof of work and you can implement this in any impact based case and this is a beautiful idea that is coming from our our our community.

The other one is impact certificates which is a revolutionary idea that enables us to not be constrained on the existing rigid structures of carbon credits, biodiversity credits that someone from the top from the United Nations created, but something that is community centered that the community can decide what is the impact that they actually want to be financed for. And hypers is a really really great standard that allows every community to basically report their impact. But of course, no one's going to trust your impact when you just report self-reported. And so that's the trick of impact certificates that allows many data sources to be connected to that certificate, right?

And one of the data sources, I think one of the coolest things you should get out from this talk is actually AI. Um, it's the sounds that you just heard. So in fact, one of the coolest things nature did about sounds is that because all the animals are screaming in the forest for mating, they are overlapping each other's frequencies and they evolved in using different frequencies for different species. So bats are using ultrasonic frequencies, frogs are using like frequencies between 0 to 1.5 kHz, birds between 1 to 5 kHz, and so they don't overlap. And even if you don't know what species that is, you can tell like from the frequency what their active is. This is like this family of

animals. And then what you can do is you can create these beautiful biodiversity sound clocks. On the left side, you see a healthy forest with colors where the maximum frequency is active. On the right side, this is a suburb. And do you notice at 6:00 a.m. and 6 p.m. these spikes? Can anyone guess what that is? >> Yes. These are the birds singing at sunrise and sunset. So you can see that in a healthy forest. You cannot you can see this also a little bit like with the frogs in the suburb of Manau. But color represents biodiversity and something that is connected to impact significant allow for more funding.

We also implemented AI that helps indigenous communities to guide. So this is a tool that helps communities report about drought and their their areas. ultimately helps them create these impact visualizations. >> So, one lesson we learned though is that when you only rely on crypto, then the crypto funding reaches only crypto people, meaning that the most people who donate when you just use crypto are crypto people from Bitcoin, right? So, we ran a Gitcoin round where we only had 46,000 uh 46 projects and only $30,000 have been raised from crypto people. Um,

but this was a credit funding round and the moment you switch it, so these are funding rounds in crypto. Round one, run two credit funding. The moment you switch it into fiat and credit card support and stripe, this is round three with a matching fund of $1 million given by Mar. Do you see how we reached out this year 11,71 individual donors because they have access they can now use basically getting incentivized by the mechanism of credit funding can use this credit card and raise more than million dollars for like projects around the world. So we believe that our mechanisms if you break them out and you allow for more payment rails we can actually enable a increase in donations.

Lastly, community first. One of the biggest principles I believe our the public goods funding committee has pioneered through Dows through governance. This is something that we strongly believe in when we collect the data is that we need to have a way for data to be governed in a fair way. So we implemented something called the nature guild which is a DAO that is running on top of the data and this DAO determines what policy this data should follow. Is it sold for pharmaceutical purposes or not? and it gets elected every 12 months copied from the public goods funding committee.

And one of the things that I really want to emphasize in this talk is that we often talk about data as the new oil or tree also mentioned that briefly like drilling data like oil and I I don't know but as a climate person I hate this. I hate the saying data is the new oil because oil means you go somewhere you extract you hurt and destroy the environment and you leave now benefit to the communities that actually own the soil right and so Audrey also mentioned we should not treat data as the new oil we should treat data as soil which is something that is means that we should tender and govern that area ourselves data is context specific and we should not have without consent extraction of the data and I think our committee is very good in creating these consents and

these governance structures to protect the data layer and the data provenence layer and ownership and at gainforce we're doing this in local workshops where communities write what can be published what cannot be published this is like a sheet life where parrots cannot shouldn't be published uh to should be published that's fine so very interesting but we can also enforce it using technology using consent records using technologies like credible exit that is enabled through cryptography so this is an exciting movement I So if you want to bring forward these ideas to other communities, I believe these ideas will survive our community and uh we have done a big big effort in working with universities like Columbia University and Oxford and ETHR to bring

these mechanisms these big ideas to the academic setting and we also gave out grants to academics to study that. If you want to learn more about this AI forpg, AI for public goods.com uh to find about our proposals and winnings and we're going to run another grants round next year as well. So lastly, I know I have four minutes left. I want to give you a really really important video that I recorded a couple years ago in the Amazon where um I just asked for the community that I worked with for a testimony of gainforest and um one of the chiefs she said something that I've never forgot that she said it's really important to not see the humans underneath the trees and I just want to share that video as a closing statement.

Florida. like so much.

So I believe the people who should look after nature should be the ones in charge and thank you so much for all the supporters over the years. Thank you so much for the public goods funding community. Without we wouldn't be here and thank you so much for the EPM community. Thank you so much. David, thank you so much for taking us from this airond conditioned and chilly conference room to the hot and humid rainforest and talk us through what you do there, the research you do and how you actually in the end preserve the rainforest. Now, there's a lot of questions from the community. Let's start right in. So, what are the biggest data gaps you currently notice in your

work? anything we collect in the Amazon and often times also in the southern Philippines are the first time anyone has ever collected data in that area like it's it's mindboggling for me as a scientist that we haven't been as humanity structured enough to finance data collection equally around the world and that is because of infrastructure gaps because of financial gaps because also of lack of incentives um to to record that area. So if you look just look at the data um UI naturalist app will tell you a lot about the birds that are living in North America. It will do okay in the Indian subcontinent but it will do horrible in in any kind of rainforest

area in the world. And I believe this is a strong data gap because this data is getting used to make decisions at the UN level who gets financed or not. And if we put crappy data into a model, you get crappy results that just emphasize an inequality. And I think that data gap is true in every level I see. Thank you. What would you answer someone who tells you AI is bad for nature? How dare you try using AI for the betterment of nature? Shout out to Emma. >> I think there is something that is very common on two things. AI has a really bad environmental footprint and we need to be very conscious about this.

Especially the large language models are really horrible in that and the data centers that we build. But there's also the concept of a handprint which is the positive effect of a model that can have to society. I give you an example on this in the Amazon that I experienced. In order to patrol and protect the rainforest, you need to go with a boat with a lot of fuel. And the Amazon is very large and it burns a lot of fuel. It's very bad for the environment. A small model with a small environmental footprint can predict which areas you should check in priority and reduce fuel fuel cost of patrols. So there's this trade-off between the handprint the positive effect of this model compared to its footprint. And we I think it needs to be a balance. There's no there's no black and white here.

>> Thank you. That's a good way of thinking about it. What new imaginations does the public goods funding ecosystem need to see? David, >> it doesn't need new imaginations. in Neo needs to reflect and do retroactively what worked really well and we have amazing ideas that work. Guys, we have created a machine. The theorem blockchain is the dream of every academic mechanism design researcher that I've ever experienced. They wanted to study their mechanisms and they need to run for government to do so. We have a machine that can basically run a smart contract and get results instantly. But they don't know about this and we don't know about them. So we need to bring these communities together in order for the public goods funding community to be taken seriously in academia but also in

academia to take seriously what we're doing here for great work and I think that's quite now lacking and that's why we started a4pg.com. We need more verification and it's good to have your topics also here on stage that we see that it makes sense to work together. Do you currently face challenges in enabling last mile payments for indigenous communities? What are things that can be made better there? it. You cannot scale faster than trust scales in the communities, right? Payments, you cannot just send someone payments and they don't don't trust you.

So, you need local champions on the ground that can explain that can build capacity that the communities trust and so you need to start small and you need to invest in these people on the ground and then that helps scale. So otherwise, it's hard to scale last mile payments. >> Thank you. Is there an advantage sonic data has over visual data? >> 100%. Like most of the animals that you go in the rainforest, they're very shy. You're not going to see them, but you can hear them because they all want to mate. They all want to meet each other. It's like a whole tinder in the rainforest happening. And so you can hear things hundreds and hundreds of meters away where you cannot even see the bird if you're not trained. So sonic acoustic data allows us to capture things that we cannot see.

Are public goodcentric chains like sal coming back? not in the current form that it was because it needs something that explains to the broader world like how can you trust these public goods projects? Where's the funding going? How do I know that I this funding is actually impactful? And so if we as a community haven't figured out that impact evation layer, donors will not come and give us money anymore. Like I think it needs to be a fair trade-off between what is your impact, how do you storyt tell that and then donors will come back and chains will come back with public goods funding. I believe so.

>> Thank you for answering these questions, David. Thank you for giving all these questions. There's a lot more questions.